PURPOSES : This study proposes a method to automatically analyze the shape of asphalt concrete aggregate using computer image analysis technology.
METHODS : To develop automatic aggregate shape analysis technology, we developed an aggregate shape boundary analysis algorithm that combines Watershed algorithm and morphology technology. After the algorithm was developed, it was mounted on a separate prototype and compared with the manual aggregate shape analysis method.
RESULTS : As a result of the algorithm comparison, the accuracy of the aggregate shape analysis was about 90%, and there was a problem that some noise occurred. This problem was derived from external roughness, the shape of the overlapped aggregate, and sand as the main causes. Therefore, we devised a method to analyze the shape of the aggregate by adjusting the external illuminance constant or using an artificial intelligence model.
CONCLUSIONS : The results of this study can be used as basic research data to help improve quality control techniques in the production of asphalt concrete in the future.

ABSTRACT
1. 서론
2. 연구배경 및 목적
    2.1. 기본 이론
    2.2. 연구목적
3. 분석방법
    3.1. 이미지 이진화
    3.2. 모폴로지 필터
    3.3. 배경영역(SureBG)의 확정
    3.4. 전경(SureFG) 내 거리 알고리즘 적용
    3.5. 불확실 영역(Unknown)
    3.6. Watershed 알고리즘 적용
4. Watershed 알고리즘 검증
    4.1. 실험 방법
    4.2. 실험 결과
5. 결론
REFERENCES

• 이강훈(한국건설기술연구원/인하대학교 토목공학과) | Lee Kanghun 교신저자
• 김연태(한국건설기술연구원/인하대학교 토목공학과) | Kim Yeontae
• 안성일(인하대학교 공과대학 스마트시티공학과) | An Sungil
• 황성도(한국건설기술연구원) | Hwang Sungdo